tcp: increment sk_drops for listeners

Goal: packets dropped by a listener are accounted for.

This adds tcp_listendrop() helper, and clears sk_drops in sk_clone_lock()
so that children do not inherit their parent drop count.

Note that we no longer increment LINUX_MIB_LISTENDROPS counter when
sending a SYNCOOKIE, since the SYN packet generated a SYNACK.
We already have a separate LINUX_MIB_SYNCOOKIESSENT

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

5 files changed, 24 insertions, 8 deletions

diff --git a/include/net/tcp.h b/include/net/tcp.h
index a23282996ca9..74d3ed5eb219 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -1836,4 +1836,17 @@ static inline void tcp_segs_in(struct tcp_sock *tp, const struct sk_buff *skb)
                 tp->data_segs_in += segs_in;
 }
 
+/*
+ * TCP listen path runs lockless.
+ * We forced "struct sock" to be const qualified to make sure
+ * we don't modify one of its field by mistake.
+ * Here, we increment sk_drops which is an atomic_t, so we can safely
+ * make sock writable again.
+ */
+static inline void tcp_listendrop(const struct sock *sk)
+{
+        atomic_inc(&((struct sock *)sk)->sk_drops);
+        NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
+}
+
 #endif  /* _TCP_H */
diff --git a/net/core/sock.c b/net/core/sock.c
index 7a6a063b28b3..2f517ea56786 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1525,6 +1525,7 @@ struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority)
                 newsk->sk_dst_cache     = NULL;
                 newsk->sk_wmem_queued   = 0;
                 newsk->sk_forward_alloc = 0;
+                atomic_set(&newsk->sk_drops, 0);
                 newsk->sk_send_head     = NULL;
                 newsk->sk_userlocks     = sk->sk_userlocks & ~SOCK_BINDPORT_LOCK;
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 0ffcd07e3409..983f04c11177 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -6339,8 +6339,10 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops,
                         inet_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
                 af_ops->send_synack(sk, dst, &fl, req,
                                     &foc, !want_cookie);
-                if (want_cookie)
-                        goto drop_and_free;
+                if (want_cookie) {
+                        reqsk_free(req);
+                        return 0;
+                }
         }
         reqsk_put(req);
         return 0;
@@ -6350,7 +6352,7 @@ drop_and_release:
 drop_and_free:
         reqsk_free(req);
 drop:
-        NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
+        tcp_listendrop(sk);
         return 0;
 }
 EXPORT_SYMBOL(tcp_conn_request);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 059a98f5e7e1..f3ce0afe70aa 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -329,7 +329,7 @@ void tcp_req_err(struct sock *sk, u32 seq, bool abort)
                  * errors returned from accept().
                  */
                 inet_csk_reqsk_queue_drop(req->rsk_listener, req);
-                NET_INC_STATS_BH(net, LINUX_MIB_LISTENDROPS);
+                tcp_listendrop(req->rsk_listener);
         }
         reqsk_put(req);
 }
@@ -1246,7 +1246,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
                                 &tcp_request_sock_ipv4_ops, sk, skb);
 
 drop:
-        NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
+        tcp_listendrop(sk);
         return 0;
 }
 EXPORT_SYMBOL(tcp_v4_conn_request);
@@ -1348,7 +1348,7 @@ exit_overflow:
 exit_nonewsk:
         dst_release(dst);
 exit:
-        NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
+        tcp_listendrop(sk);
         return NULL;
 put_and_exit:
         inet_csk_prepare_forced_close(newsk);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 5fa8fea394c9..7cde1b6fdda3 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -964,7 +964,7 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
                                 &tcp_request_sock_ipv6_ops, sk, skb);
 
 drop:
-        NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
+        tcp_listendrop(sk);
         return 0; /* don't send reset */
 }
 
@@ -1169,7 +1169,7 @@ out_overflow:
 out_nonewsk:
         dst_release(dst);
 out:
-        NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
+        tcp_listendrop(sk);
         return NULL;
 }